AI is a bubble, but it won’t burst anytime soon, says Viktor Shvetz

Artificial Intelligence

• AI bubble aids adoption, unlikely to cause prolonged market crash
• AI investments soar, but end-user revenue and productivity gains are limited.
• India's stock markets might be undervaluing long-term AI-driven risks.

Did our AI summary help?

AI, like previous waves of transformative technology, is running on a bubble—but one unlikely to trigger a prolonged market crash. Speaking at the CFA Institute India Investment Conference, Viktor Shvetz, global strategist and author, said every general-purpose technology (from railways to telecoms) requires a capital-fueled bubble to drive adoption and scale.

“We have abundant capital, not shortage… If you have abundant capital, that not only cushions the economies, but it also cushions the market,” he explained.

While AI valuations could experience sharp, short-term declines, such corrections are likely to be much shallower and shorter than past tech bubbles. “It could be something like happens on Thursday and things went to waste… the recovery stage will be much shorter,” Shvetz said.

AI remains early in its technology cycle. Investments, particularly in hyperscale infrastructure, are already massive, but end-user revenue remains limited. According to Shvetz, the market has not yet reached the kind of peak seen during telecoms in the late 1990s.

In his view, the AI bubble is both inevitable and necessary, allowing the technology to permeate infrastructure, applications, and industrial processes, laying the foundation for long-term societal and economic transformation.

Shvetz also highlighted the changing global innovation landscape. While the US historically combined inventiveness with innovation, he warned that recent policy decisions are undermining its ecosystem. “The United States over the last four months was shooting itself in the foot, both in terms of clusters and in terms of funding our fundamental research. That’s bad news,” he said.

China, by contrast, has excelled at innovation even without deep inventiveness. “Over the last 40 years, China invented absolutely nothing. But it has been the world’s greatest ever innovator of products and services,” Shvetz said. He added that machines increasingly drive innovation, reducing the necessity of personal freedom: “That is exactly the point Xi Jinping is making. You can actually create those clusters without ceding a great deal of control.”

Despite this, structural risks remain. China faces massive overinvestment and misallocated capital, while both China and the US aim for the elusive goal of 5% productivity growth with zero inflation, which could take a decade or more to achieve.

Shvetz warned that markets may be underestimating the implications of AI on labor, capital, and social stability. “Output per employee is rising, but multi-factor productivity has not taken off… We are at least a decade away from that,” he said. He also cautioned about the social consequences, including rising inequalities and labor polarization over the next 10–15 years.

Turning to India, Shvetz said equity markets are underpricing long-term structural risks. “Equity fund managers can only value what they can value. They don’t look at things they cannot value, and those things come back to bite you,” he said. Without policies to manage inequalities and support skilling, the benefits of AI-driven growth could remain concentrated, leaving large segments of the workforce underrepresented in prosperity.